ups limit?
-
- Posts: 4022
- Joined: Sat Mar 12, 2005 6:24 pm
-
- Posts: 4022
- Joined: Sat Mar 12, 2005 6:24 pm
Keeping in mind that Q3A uses vectors represented by three floats[1], you could - if two of the XYZ components remain small[2] - reach much higher speeds. A vector's length equals the square root of X^2 + Y^2 + Z^2. Now let us assume Y and Z stay within a 0-200 range so the sum of their squares never exceeds 200^200 + 200^200 = 80.000. This leaves us with an X component that can be as high as 2^127 - (80 * 10^3). After calculating the square root we have a velocity of approx. 2.89 * 10^76 units per second[3]. Yes kids, that's a two followed by seventy-six zeroes.
[1] To be more specific: 32 bits, single-precision, IEEE 754 style floats that have an upper limit of about 2^127 or about 1.7 * 10^38.
[2] For Z, usually 0 <= Z <= 200 will hold true. X or Y can be kept low if you move along the east/west or north/south axis.
[3] Unfortunately, Q3A uses three floats to store the location as well, so this would overflow. More over, floats loose a lot of precision when they become > 2^40.
[1] To be more specific: 32 bits, single-precision, IEEE 754 style floats that have an upper limit of about 2^127 or about 1.7 * 10^38.
[2] For Z, usually 0 <= Z <= 200 will hold true. X or Y can be kept low if you move along the east/west or north/south axis.
[3] Unfortunately, Q3A uses three floats to store the location as well, so this would overflow. More over, floats loose a lot of precision when they become > 2^40.
-
- Posts: 4022
- Joined: Sat Mar 12, 2005 6:24 pm